Detection of brain death onset using the bispectral index in severely comatose patients

To evaluate the accuracy of bispectral index (BIS) monitoring for the diagnosis of brain death in severely comatose patients.

A prospective study in an intensive care unit of a university hospital.

Fifty-six severely comatose patients (Glasgow Coma Score < or = 5) admitted to the ICU mainly because of intracerebral hemorrhage, head injury, or postanoxic coma.

BIS was recorded continuously during the hospitalization in the ICU. Where necessary, clinical brain death was confirmed by EEG or cerebral angiography.

Twelve patients were already clinically brain dead at the time of admission, and their individual BIS values were 0. In each of these 12 patients brain death was thereafter confirmed by EEG or cerebral angiography. Forty-four patients were not clinically brain-dead at the time of admission, and their individual BIS values were between 20 and 79. Twenty-seven of these patients became brain-dead, and their individual BIS values dropped to 0 in a few hours to a few days. In these 27 patients EEG or cerebral angiography was performed after the BIS value decreased to 0 and confirmed brain death in all cases. Seventeen patients who did not become brain dead during their hospitalization in the ICU had persistent electrocerebral activity on EEG, and their average BIS values remained above 35.

BIS can be used in severely comatose patients as an assessment of brain death onset, enabling appropriate scheduling of either EEG or cerebral angiography to confirm brain death.